Although it is well-established that schedule-induced drinking occurs on intermittent schedules of food reinforcement, the mechanism by which drinking arises is still a mystery. The proposed experiments will test four major hypotheses. First drinking is a conditioned response that is controlled by the same basic laws as other conditioned responses. Second, drinking arises in anticipation of upcoming food delivery rather than occurring as a reaction to the previous pellet. Third, competition occurs between drinking and the operant, which can result in the displacement of drinking to the early portion of the interpellent interval. Finally, drinking may be controlled by timing mechanisms in much the same way as operant responses. The experimental results will be interpreted with three major theories of timing to determine whether the temporal locus and duration of bouts of drinking may be theoretically modeled in the same manner as operant responses such as lever pressing. If necessary, modifications in the theories will be made to accommodate the results of the proposed experiments.